Research On Channel Estimation Of Mm Wave Massive MIMO System Based On Low-Rank Matrix Recovery

As the core technologies of 5G and 6G,massive Multiple Input Multiple Output(MIMO)and mm Wave technology complement each other,which further expand the antenna scale and improve the system performance.However,the system becomes more complex,and the traditional channel estimation method will lead to excessive pilot overhead.Therefore,we use the sparse characteristics of mm Wave’s angle domain to solve the channel estimation by the method of low-rank matrix recovery.Therefore,the thesis studies channel estimation algorithm based on the joint weighted and truncated nuclear norm and the channel estimation algorithm based on Parallel Factor(PARAFAC)constrained alternating least squares,aiming at reducing the training overhead of communication system,reducing the computational complexity and improving the estimation accuracy of Channel State Information(CSI).The main research content is as follows:Firstly,the thesis studies channel estimation algorithm based on low-rank matrix recovery,the joint weighted and truncated nuclear norm is used to approximate the relaxation of rank function,a new matrix recovery model is constructed for channel estimation.The optimization goal is to minimize the weighted and truncated nuclear norm,and the alternative optimization framework is used to solve it.The simulation results show that this method effectively improve the accuracy of channel estimation and has reliable convergence.Secondly,in order to improve the accuracy and convergence speed of estimation of massive MIMO cascaded channels assisted by Intelligent reflective surface(IRS),based on the Parallel factor(PARAFAC)decomposition model,the thesis studies a constrained bilinear alternating least squares(CBALS)channel estimation for the problems of slow convergence speed and long iteration time in the traditional ALS algorithm.By using the asymptotic orthogonality of massive MIMO channel matrix,we integrate it into the iterative process of fitting algorithm as a constraint.The simulation results show that compared with the existing pilot-based channel estimation algorithms,the proposed algorithm can significantly improve the accuracy of channel estimation,and the proposed CBALS algorithm has better estimation performance,faster convergence speed and lower complexity in the case of high power.